M3S-GRPred: a novel ensemble learning approach for the interpretable prediction of glucocorticoid receptor antagonists using a multi-step stacking strategy

Table 2 Cross-validation MCC scores of 30 base-classifiers developed based on individual five balanced training datasets

Descriptor	Method	BTS1	BTS2	BTS3	BTS4	BTS5	Average
AP2DC	KNN	0.509	0.342	0.446	0.441	0.470	0.442
	MLP	0.508	0.449	0.513	0.472	0.507	0.490
	PLS	0.498	0.425	0.404	0.398	0.501	0.445
	RF	0.482	0.487	0.513	0.455	0.513	0.490
	SVM	0.549	0.524	0.548	0.459	0.526	0.521
	XGB	0.540	0.494	0.533	0.497	0.457	0.504
CDKExt	KNN	0.469	0.421	0.431	0.400	0.487	0.442
	MLP	0.426	0.477	0.484	0.443	0.496	0.465
	PLS	0.365	0.474	0.464	0.478	0.472	0.451
	RF	0.450	0.519	0.494	0.522	0.521	0.501
	SVM	0.517	0.555	0.489	0.476	0.516	0.510
	XGB	0.483	0.479	0.444	0.472	0.516	0.479
FP4C	KNN	0.417	0.382	0.451	0.395	0.447	0.418
	MLP	0.007	0.031	0.050	-0.018	0.111	0.036
	PLS	0.436	0.434	0.400	0.341	0.452	0.413
	RF	0.478	0.509	0.519	0.480	0.545	0.506
	SVM	0.468	0.500	0.548	0.454	0.545	0.503
	XGB	0.520	0.539	0.503	0.495	0.560	0.524
MACCS	KNN	0.323	0.392	0.395	0.337	0.410	0.371
	MLP	0.017	0.040	-0.050	-0.023	-0.058	-0.015
	PLS	0.273	0.336	0.335	0.267	0.394	0.321
	RF	0.406	0.524	0.539	0.440	0.527	0.487
	SVM	0.339	0.510	0.493	0.405	0.443	0.438
	XGB	0.426	0.494	0.514	0.478	0.472	0.477
Pubchem	KNN	0.442	0.379	0.374	0.452	0.441	0.417
	MLP	0.468	0.469	0.404	0.546	0.511	0.480
	PLS	0.554	0.505	0.444	0.526	0.516	0.509
	RF	0.563	0.509	0.499	0.528	0.531	0.526
	SVM	0.491	0.460	0.451	0.563	0.526	0.498
	XGB	0.540	0.433	0.464	0.546	0.506	0.498

ISSN: 1471-2105